THE INTEGRATION OF ARTIFICIAL INTELLIGENCE AI IN TEACHING SCIENCE AMONG PUBLIC SCHOOLS IN THE PHILIPPINES BEST PRACTICES, CHALLENGES, AND READINESS OF TEACHERS Lovely Joy Abcede-Morales, MAED, Benjie P. Jacob, PhD Graduate School, Saint Joseph Institute of Technology, Butuan City, Agusan Del Norte, Philippines https://doi.org/10.5281/zenodo.15767527 ABSTRACT This study determined the best practices of AI integration in ASEAN countries and leading public schools in the Philippines, the challenges faced by science teachers in integrating AI, and the readiness of science teachers to integrate AI in teaching science. It uses a multi-method approach, including a systematic literature review, focus group discussions, interviews, and a quantitative survey. The SLR method provides a rigorous and transparent process for identifying evidence-based practices. The qualitative phase gathers insights from science teachers' experiences, perceptions, and challenges. Focus group discussions and interviews with the three largest public schools in Butuan City collect in-depth data. The quantitative survey assesses the readiness of science teachers to adopt AI in their teaching practices, focusing on technology self-efficacy, student interaction, and ethical awareness, using an adapted survey questionnaire of Ramazanoglu and Akın (2025) AI readiness scale for teachers. Findings revealed that AI integration in science teaching using various tools like ChatGPT, Canva, Google Gemini, Google Classroom, Microsoft Copilot, Grammarly, QuillBot, and Turnitin are the best practices in ASEAN countries and prominent public schools in the Philippines. Science teachers face challenges in integrating AI tools due to high costs, content quality, and relevance. They find AI-generated content irrelevant, decrease teacher-student interaction, and face creativity blocks. Access to technology and infrastructure is difficult, and ethical concerns include academic dishonesty, data privacy violations, and student unawareness. Overcoming these challenges is crucial for effective AI integration in